The Toll pathway mediates Drosophila resilience to Aspergillus mycotoxins through specific Bomanins

Article activity feed

1. 

   Review Commons

   Nov 17, 2022

   Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Reply to the reviewers

   Reviewer #1 (Evidence, reproducibility and clarity):

   Summary
   The authors have set out to study the Drosophila immune response against the fungus Aspergillus fumigatus. They found that Aspergillus fumigatus kills Drosophila Toll pathway mutants. The fungus does this without invasion because its dissemination is blocked by melanization. They suggest that there is a role for Toll in host defense distinct from resistance. The findings are interesting, and looks like the mycotoxins play a role. It also seems that there is some role of the Bomanins here, but I find that in particular Figure4 experiments are not convincing enough to provide a mechanistic insight …

   Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Reply to the reviewers

   Reviewer #1 (Evidence, reproducibility and clarity):

   Summary
   The authors have set out to study the Drosophila immune response against the fungus Aspergillus fumigatus. They found that Aspergillus fumigatus kills Drosophila Toll pathway mutants. The fungus does this without invasion because its dissemination is blocked by melanization. They suggest that there is a role for Toll in host defense distinct from resistance. The findings are interesting, and looks like the mycotoxins play a role. It also seems that there is some role of the Bomanins here, but I find that in particular Figure4 experiments are not convincing enough to provide a mechanistic insight as to what is going on. I think the authors need to think through what their results mean, and also, explain better (especially regarding Fig 4) their ideas and how the data fits them.

   We thank the reviewer for scrutinizing our manuscript as well as for suggestions to improve it.

   The role of mycotoxins is demonstrated:

   i) the fungus does not proliferate nor disseminate, also in Toll pathway mutant flies: thus, it must kill through diffusible substances, in as much as these immuno-deficient flies exhibit tremors toward the end of the infection;

   ii) a fungal strain devoid of the capacity to produce secondary metabolites is no longer virulent, even in Toll pathway mutant flies.

   The role of Bomanins is also demonstrated: the finding of a susceptibility of Bom__D__55C deletion flies to A. fumigatus and to mycotoxin challenges clearly shows that at least one or several Bomanin genes are required in the host defense against these challenges. The observation that this susceptibility can be rescued by the genetic overexpression of specific Bomanins indicates which ones are likely to mediate protection. The novel data we have included with the protection from mycotoxin action in neurons point clearly to BomS6 being the major mediator of protection against verruculogen action since it is the only one of two Bom genes to be induced in the head and with a proven potential for rescue of the Bom__D__55C phenotype.

   As regards the concept of the article, it is simple: we show that the Toll pathway does not control A. fumigatus infection by directly attacking the fungus but does so by neutralizing the effects of secreted virulence factors such as restrictocin and verruculogen. We further identify some of the relevant effectors such as Bomanins by using a genetic complementation strategy. To make our point clearer, we have now included additional data in which we show that BomS6 and BomS4 are the only Bomanins induced in the head of flies upon the injection of these two toxins. We next determine that BomS6 and not BomS4 expression in the nervous system dominantly protects the flies from the deleterious effects of verruculogen injection, both in terms of recovery from tremors and survival. Mechanistically, the Toll pathway protects the host from the action of verruculogen by expressing and likely secreting BomS6 from neurons.

   Major comments:
   Page 5: .."the fungal burden did not increase much in MyD88 flies challenged with 50 conidia (Fig. 1B)" - What do you mean did not increase much? There is a clear increase in Myd88 mutants compared to controls; would you expect a bigger increase (e.g. log scale induction)? Explain.

   When the injected dose is higher than 50 injected colonies, the fungal burden remains very close to that of the injected inoculum (Fig. EV1_F, J_). As for other pathogens regulated by the Toll pathway, it has been published that the microbial burden increases by log factors for filamentous fungi (Huang et al.., in revision), pathogenic yeasts (e.g., work from our laboratory Quintin et al. Journal of Immunology, 2013), bacteria (e. g., Duneau et al., eLife 2017; Huang et al., in revision). The pathogens usually proliferate exponentially in immuno-deficient hosts, which is clearly not the case of A. fumigatus, the first example we know of.

   Page 6: "the SPZ/Toll/MyD88 cassette is required for host defense against A. fumigatus infections, even though this pathogen only mildly stimulates the Toll pathway." - Should you rather say that A. fumigatus only mildly induces the Toll pathway target gene Drosomycin?

   The answer is negative. Fig. EV1_C_ clearly shows that BomS1 is also modestly induced as compared to an infection with E. faecalis. The promoter of BomS1 contains a canonical Dif-response element (Busse et al., EMBO J., 2007_)_. For a more thorough discussion of this point, please, see reply to Reviewer 2, Major Comment 2.

   Page 6: "...we tested Hayan mutant flies defective for this arm of innate immunity (Nam et al., 2012)." - elaborate this, which arm/which pathway?

   The title of the paragraph is “Drosophila melanization curbs A. fumigatus invasion”. The full first sentence of the paragraph actually read: “As melanization is a host defense of insects effective against fungal infections, we tested Hayan mutant flies defective for this arm of innate immunity”.

   This has not been introduced in the introduction. Explain.

   We have now added a couple of lines (82-83) to introduce melanization for the nonspecialist reader.

   Can you really draw this conclusion: "We conclude that melanization limits the proliferation and the dissemination of A. fumigatus injected into wild-type flies yet does not eradicate it at the injection site, where a melanization plug forms." Maybe you can based on the function/importance of the pathway to melanization, but you need to explain.

   Melanization is mediated by the Hayan protease and three phenol oxidases (two in adults) that catalyze the enzymatic reactions leading to melanin production (for Drosophila, please see Nam et al. EMBO J. (2012), Bingelli et al., PLoS Pathogen (2014), Dudzic et al., BMC Biology (2015), Cell Reports, 2019). Thus, finding that there is an increased proliferation and dissemination in null Hayan mutants is a strong indication for a role of melanization. The identification of a similar phenotype for PPO2 and PO1-PPO2 mutants demonstrates that melanization is curbing A. fumigatus. Our sentence is therefore fully justified.

   Page 10: "The cleavage of the 18S RNA was however much less pronounced in wild-type flies as compared to MyD88" - I am not sure what this means. Do you mean 28S?

   We thank the reviewer for pointing out this mistake that has now been corrected.

   And that the 28S peak is lower? Is this a quantitative method?

   The technique is liquid electrophoresis on a microchip. It is both a qualitative and quantitative technique that replaces traditional agarose or polyacrylamide gels.

   Fig. legend: "Arrows show the position of the 28S RNA sarcin fragment" - there are three arrows in both Fig 4E and F; specify which arrows point what.

   The thick arrow is now indicated in the figure legend to correspond to the much smaller sarcin fragment whereas the thin arrows on the graph clearly specify the position of the 28S RNA peaks.

   Based on the results, I am not convinced about the conclusion, that "restrictocin is able to inhibit translation to a detectable degree in vivo, likely through the cleavage of the ribosomal 28S a-sarcin/ricin loop as described in vitro." <- Do you draw this conclusion before doing the actual in vitro experiment, which is described next in the text (The rabbit reticulocute assay, S2 cells)?

   The existing literature (line 259 for a few selected references) has largely proven that restrictocin cleaves 28S RNA in vitro. We are demonstrating that this also happens in vivo in flies based on the generation of the alpha-sarcin fragment as well as the decreased 28S peaks. Our transgenic approach also indicates that restrictocin blocks translation in vivo. The in vitro approach has been implemented so that we could test the effect of synthetic BomS1 and BomS3 in cell culture. As to our knowledge, no one had demonstrated that restrictocin blocks translation in Drosophila cultured cells. It was therefore important to demonstrate it in cell culture using well-characterized in vitro techniques mastered by AT and FM.

   4H: Not sure what should be seen here, is it the darkest band at 0 uM that disappears?

   We have improved the figure and added an arrow to point out to the relevant band on the gel.

   HI & J need more explanation than what is now included in the text or Figure legend, is the conclusion that there is no difference? Write the stats above the Figs 4I & J (n.s.?).

   We have added NS on the figures and made our conclusion clearer (lines 295-298).

   Minor comments:

   It would have helped commenting if the manuscript contained line numbers

   We apologize for having initially provided a version in which lines were not numbered. At the prompting of Review Commons we immediately provided such a version, that was actually used by Reviewer 2.

   Why do you have the title "Hayan" on top of Fig 1F; you don't have this marking system in the other survival curves

   This point has now been addressed and the survival experiments checked for consistency.

   Fig 2A: Can you speculate why MyD88 flies die rapidly at day 10 if you inject PBST (your control)? What would happen to uninjected controls in otherwise the same conditions? (you could include an uninjected control here?)

   We suspect that this is linked to the trauma induced by the injection. Trauma has been shown to impact the homeostasis of the midgut epithelium (Lee & Miura, Current Topics Developmental Biology 2014, Chakrabarti et al., PLoS Genetics (2016)), and we suspect that it may lead to a leakiness of the gut allowing the passage of some bacteria from the gut microbiota that can proliferate in the hemocoel. Hence, we checked axenic and antibiotics-treated MyD88 flies to exclude that the limited sensitivity to trauma was not significantly contributing to the phenotypes we describe. It is also linked to the thickness of the needle and the problem is alleviated by using thinner needles.

   The uninjected control is now shown in Fig. EV8_E_.

   Please, see also the answer to Reviewer 2 Major comment 1.

   Fig 2E: Not sure what would be the best way of presenting the curves - different colors, dotted lines or something? Now if there are too many lines, they are hard to tell apart. because the symbols are not that visible. Like in 2E if you want to compare the light red/orange colored lines.

   We agree with the reviewer that the lines are hard to tell apart. This is however not a significant issue since the glip mutants display curves similar to that of the wt A. fumigatus control strain.

   For consistency add the caption also to Fig 3D (I assume it is the same as 3C)

   The caption was present in our version and is present in the revised version.

   For consistency, should you add Verruculogen on top of Fig 3F?

   Same reply as for the previous comment.

   Chronologically, how it is explained in the text, Figs 4A and B are in the wrong order.

   We fully agree with the reviewer. This problem has been addressed in the revised version.

   The quality of Fig 4 is not great, the text is hard to read (too small) and becomes blurry upon magnification.

   We fully agree with the reviewer. This problem has been addressed in the revised version.

   Page 12; "These data then suggest that a process akin to the immune surveillance of core cellular processes first described in C. elegans may also exist in Drosophila" - I think this sentence belongs to the discussion, this is not directly drawn from the results.

   We have followed the reviewer suggestion and have now developed our Discussion paragraph now entitled “Induction of the expression of specific Bomanin genes upon mycotoxin challenge”

   Referees cross-commenting

   I think we share many thoughts among all the reviewers.

   The main problem is that the manuscript language is quite strong; from the results many times it is not ok to make such strong statements. Some experiments need further analysis and clarification.

   I think in most cases, this could be achieved by softening the statements and adding more discussion, and not by making new experiments (some may be needed).

   We respectfully disagree with the reviewer on this point. There were obviously some misunderstandings that might be traced to the short format of the initial version. We have now developed the Discussion to clarify our conclusions as suggested by the reviewer.

   Minor things are that experiments are not advancing in a logical order between the text and the figures and there are problems with resolution in some figures.

   Statistics in some figures needs to be added.

   Please, see above.

   Reviewer #1 (Significance):

   The nature of the work is conceptual for the field, to understand the role of the Toll pathway and Bomanins in particular, in this fungal infection model. The work is interesting to a somewhat limited audience, mainly immunologists and in particular, people interested in the Drosophila model for immunity. The work may be interesting conceptually in understanding fungal infections.

   We are not certain that immunologists represent a limited audience. We agree that work on fungal infections is insufficiently funded with respect to the medical importance of these infections, as highlighted in our introduction and Perspective section of the Discussion.

   My expertise: I am a Drosophila immunity researcher with nearly 20 years of experience in working with fly immunity, in particular the Toll and the Imd pathways.

   Reviewer #2 (Evidence, reproducibility and clarity):

   Summary:

   Xu et al. describe how A. fumigatus kills Toll-deficient fruit flies not by hyperproliferation, but more likely by virulence factors. Melanization is important for suppressing fungal spread. The Bomanin genes have an unknown function, and here the data suggest a reasonably convincing role for Toll in resilience. Overall the manuscript is thorough and presents a diversity of approaches that show Toll and the Bomanins in particular contribute to this resilience effect. The idea that Toll effectors are essential for resilience is interesting as other fly stress response pathways like JAK-STAT are better known for helping the fly cope with damages, while Toll is better known as an antifungal response.

   I believe the study, with some careful considerations added, would add a valuable series of observations to understanding how the host immune system promotes survival after infection. Overall I am quite positive about the results, and the authors have made a significant effort.

   We thank the reviewer for the positive evaluation of our work that actually spans many years of research on the Aspergillus fumigatus Drosophila infection model that is a major topic of our work at the Sino-French Hoffmann Institute of Guangzhou Medical University.

   Any experiment suggestions I make are strictly to improve the confidence in the interpretations of the results, but the language could alternately be softened to address those concerns. My major critique is that the authors repeatedly extend beyond what is shown, and occasionally in defiance of what is shown (if I understand the results correctly).

   We have chosen to perform additional experiments when needed. We have also clarified points where there were obvious misunderstandings by expanding our text that had been written under a very concise format.

   It is not thoroughly clear what the reviewer has in mind when using the word defiance. We suppose it refers to the work of Scott Lindsay with whom we are in contact. He actually attempted to monitor the C. glabrata burden but did not pursue this line of investigations as he already saw a difference after one hour and he thought that the Toll pathway cannot be induced so rapidly. Actually, David Duneau mentions a time of two to three hours for the Toll pathway to control E. faecalis infections (eLife, 2017) and Sandrine Uttenweiler-Joseph already saw by MALDI-TOF MS an induction of Bomanins and other DIMs at the earliest point tested, six hours (PhD thesis). There is absolutely no critique of the work of the Wasserman laboratory who has greatly contributed to our understanding of Bomanin functions. Some of our unpublished data clearly point out to an AMP role for at least one Bomanin gene against E. faecalis and we certainly do not exclude an AMP role for BomS against C. glabrata. This however does not dismiss the possibility that Bomanins may also have other roles in dealing with microbial toxins. We have been studying Candida infections in Drosophila for many years and have documented the host defense against C. glabrata (Quintin et al., JI, 2013). We do suspect that C. glabrata likely secretes virulence factors that have not been identified so far. We mention this as a possibility and certainly not as a truth. One should remember that investigators were unaware for a long period of the role of Candidalysin, a pore-forming toxin, in C. albicans infections.

   Finally, a dual role as AMP and protecting from secreted toxins has been clearly shown in the case of alpha-mammalian Defensins that we now are describing in our Revised Discussion (Kudryashova,Immunity, 2014).

   Comments below.

   Major comments:

   1. The language is too strong. Specifically the use of the phrase "anti-toxin" is too generalist, especially as the authors show that their candidate Bomanin does not bind to the toxin directly.

   We have checked all of the submitted documents: the term anti-toxin was never used (just found “anti” in antimicrobial, antifungal, antibiotics..), in this manuscript as well as in the companion article. and we have never excluded an indirect effect, quite to the contrary because of the in vitro experiment with restrictocin mentioned by the reviewer and other observations now included (see further below). We use the terms “protection” or “counteract”, which have not such a meaning. It is burdensome for the reader to read each time “counteract or protect from the actions of the toxins or the effects of the toxin.

   Instead, Toll mutants seem susceptible to damage/stress caused by injury/toxins. MyD88 even show general susceptibility to vehicle controls in Fig3C-D.

   The effects of stress related to the infection conditions and injury are clearly distinct from the much stronger ones exerted by the toxins themselves. As requested by the reviewer further below, we have submitted wild-type and immuno-deficient flies to several stresses such as heat or the injection of hydrogen peroxide or salt solution (Fig. EV8_B-E_). While the latter did not reveal any difference, MyD88 flies succumbed slightly faster to a strong 37°C stress; in contrast, they survived better to a 29°C exposure, the temperature at which we perform most experiments. However, the difference started to be visible only after some 15 days whereas the time frame in which flies succumb to A. fumigatus or toxin challenges is definitely much shorter by some 10 days. We also note that Bom__D__55C mutant flies behave like the isogenized wild-type controls in these assays, further excluding a potential role for general stress sensitivity as a contributor to the effect of toxins.

   As regards DMSO, there is indeed a general mild sensitivity of flies to DMSO, but not specifically affecting MyD88 mutant (Rebuttal Fig. 1J). We find that this effect is lessened when using thinner needles. Thus, the problem has become minor as we became more experienced. We had checked axenics- and antibiotics-treated flies to exclude a contribution from the microbiota. Finally, to uncouple the effects of verruculogen from those of DMSO, we have also challenged flies directly by introducing the powder, using a technique similar to that of the septic injury. While it is quantitatively less accurate, it clearly proves that verruculogen produces the reported effects (Fig. 3C) and was useful to measure Bom and Drosomycin expression by digital PCR in the heads of challenged flies, e.g., Fig. EV6_J-K_ and Figs EV_11&12_.

   Toll is important for development, so it may be expected that Toll flies could have development defects impacting resilience even if/when Toll flies can survive to adulthood. I don't say this to be too negative on the findings, which are quite convincing. But I am not sure that the phrase "anti-toxin" is right for what is shown.

   We fully agree with the reviewer on this point. We have failed to find RNAi lines that are efficient enough to mimic the Toll pathway phenotype when expressed ubiquitously at the adult stage. However, Bom__D__55C mutants do not seem to display a developmental phenotype and display a phenotype similar to that of MyD88 flies. Furthermore, our rescue experiments of the Bom__D__55C sensitivity phenotype to mycotoxin challenge is achieved by the overexpression of specific Boms that are induced only at the adult stage, making it unlikely that this sensitivity phenotype reflects a developmental problem, as had been shown to be the case for 18-wheeler that had initially been proposed to encode the IMD pathway receptor.

   A very interesting recent study shows Dif has a role in the synapse of neurons to protect from alcohol sensitivity. Could secreted Bomanins participate? This emphasizes a mechanism through which Toll mutants likely have defective neural development, which could make them stress response defective, especially to things like neurotoxins. See: https://pubmed.ncbi.nlm.nih.gov/35273084/

   We are aware of this study first presented at the 2019 Fly Meeting in Dallas and this author did discuss with the authors of the study. However, we have found that Dif (and Dorsal) mutants are not sensitive to A. fumigatus infections nor to injected mycotoxins, as was the case already for C. glabrata (Quintin et al., JI, 2013).

   Lin et al. (2019) also showed lack of Bomanin secretion from the fat body in Bombardier mutants causes loss of tolerance (resilience?). So does Bomanin disruption increase susceptibility to stresses more generally, rather than specifically fungal toxins? And is this a development role, rather than an immune response role?

   The authors could try to use other stresses (NaCl, oxygen, heat, alcohol) to test the contribution of Bomanins to this resilience, which may reflect defective neural development rather than a role for secreted systemic immune-response peptides.

   Please, see replies above.

   1. The authors present a paradox. On the one hand, A. fumigatus hardly induces Drs/Bomanins (Fig. S1). Yet on the other, they propose that inducible Bomanins protect the fly from mycotoxins. Why do the authors say Toll is hardly induced by A. fumigatus at the start of the study (Fig S1), but later use the same data to argue that Bomanin induction underlies the resilience phenotype (Fig5).

   The reviewer raises an interesting point. Of note, we have added new data in Fig. EV2_B_ that document that all 55C Bomanin genes, _BomS4-_excepted, are induced by a systemic infection. There is indeed somewhat of a paradox. The Bom__D__55C deletion phenotype clearly establishes that Bomanins play a major role in the protection against mycotoxins and A. fumigatus. The rescue experiments rely on ectopic expression and therefore establish that specific Bomanins can mediate the protective effect. Our data on verruculogen suggest that there might be local inductions, e. g., in the head of BomS6 and BomS4. The brain represents a compartment that is separated from the hemocoel by the blood-brain-barrier. We have not been able to generate BomS6 null mutants so far. In this case, the relevant response may not be systemic. We only detect a weak signal for BomS peptides in the hemolymph of unchallenged flies, making it unlikely that a basal expression is important, at least as regards a systemic infection. We cannot however exclude local inductions at the level of tissues. This would not rely on hemocytes as “hemoless” flies are not susceptible to A. fumigatus or toxin challenges. This topic definitely warrants further investigations.

   In Fig 5, it looks like DMSO is nearly identical to A. fumigatus, so can the authors really suggest that equal induction to DMSO is relevant?

   We had stated that an induction of the Bomanins by the injection of DMSO alone precluded us from analyzing the effects of verruculogen on Bom gene expression. We have now bypassed this difficulty through direct challenges by the undissolved powder (Fig. 6_J-K,_ Fig. EV11).

   The authors' discussion of these points would benefit from considering Vaz et al. (2019; Cell Rep) to frame how much PAMP is injected given equal numbers of fungal cells vs. bacterial cells. To me the lower induction by injecting a few fungal cells with much lower surface area to volume ratio means equal microbe mass has exponentially less PAMP in fungal conidia cell walls (2-3um diameter) vs. equal mass of bacteria (0.5-1um diameter).

   We fully agree with the reviewer and now mention that C. glabrata also led to a milder induction of the Toll-mediated humoral response (Quintin et al. JI, 2013). In addition, it has been shown previously that ß-(1-3)-glucans, which are sensed by GNBP3 in Drosophila (Gottar et al., Cell, 2006), are concealed by the cell wall (germinating conidia) or hydrophobins (Wheeler et al., PLoS Pathogens, 2006; Aimanianda et al., Nature, 2009) . In the case of yeasts, these glucans are accessible only at the budding scar (Gantner et al., EMBO J., 2003).

   Fig S1O is not convincing that Boms alone are present. There is significant noise near Drs in FigS1 infected, which likely saturates the detector before Drs can fly to it. I say this because DIM4 (Daisho) indicates that Toll is strongly induced. The authors should show a larger mass range on the x-axis including peaks of other Toll-induced peptides like the BaramicinA DIM10, DIM12 and DIM13 peptides of their companion paper and DIM14 (Daisho), which are closer in mass to the Bomanins and less likely disrupted by the noise at 4300 m/z. The maldi-tof calibration to correct ranges is critical for arguments of quantification.

   We provide the primary data in the Rebuttal figures at the end of this document. These are the results obtained from three single flies (Files A29683PBUG22, A29684PBUG23 and A29684PBUG24). The first three spectra correspond to the full scale based on the major peaks observed (DIM4/BomS5) in two out of three spectra. At this scale, no signal is visible for Drosomycin at 4891 and the “noise” at 4278 is modest. Next, the multi-spectra report allows to put all three samples on the same sheet, this time zooming on the peaks of interests in the region 4300 (“noise”) and 4891 (Drosomycin). Finally, the next two pages zoom in on the BomS peptide signals and the next page keeps the same scale to document the 4300-5000 region. On the last page, it is obvious that the signal around 4300 is very modest and too distant to influence the Drosomycin ion, thereby excluding any effect of suppression. Of note, in the systemic immune response, Drosomycin is the most induced AMP with a concentration estimated to be around 0.3µM, an order of magnitude higher than other AMPs. Finally, these experiments have been performed by PB who initially developed the technique (Uttenweiler-Joseph, PNAS, 1998) and has been using and developing it ever since.

   Combined with comments in Major Concern 1, I am not convinced that the -inducible- Bomanin response mediates the resilience phenotype.

   Besides our replies above, we do hope that the new data we have included in Fig. 6 that document an induction of only two BomS genes in the heads of Drosophila upon verruculogen and the finding that BomS6 expression in the nervous system protects the fly from the effects of verruculogen will convince this reviewer.

   1. The author's language is very strong to disregard a possible antimicrobial activity.

   As noted above, this is a misunderstanding that we hope is dispelled in the revised discussion (see also above and replies to Reviewer 1).

   Previous studies showed increased Candida growth and decreased hemolymph killing activity in Bom55C flies (Lindsay et al. 2018 and Hanson et al. 2019).

   Please, see reply above. Factually, Lindsay et al. did not study the C. glabrata titer in vivo but using collected hemolymph. The killing activity likely requires a cofactor regulated by the Toll pathway. Hanson investigated the burden of the dimorphic C. albicans pathogen that in flies is filamentous and not C. glabrata.

   Also see minor concern (i).
   I grant that the data are consistent with a resilience role. However the authors found no binding of Bomanin to restrictocin, countering their idea of a -direct- anti-toxin effect.

   We are surprised by this comment. We certainly did not favor this idea nor developed it in the original manuscript, even though we cannot formally exclude it at this stage. Future experiments will focus on BomS6 potential interactions with these two mycotoxins.

   At present the authors cannot rule out a direct antimicrobial role, or even the possibility of two different roles for the same peptides (ex: one in resilience, one antimicrobial). For instance, it is difficult to explain the loss of killing activity of Bom-deficient hemolymph ex vivo from Lindsay et al. if Bomanins are strictly anti-toxins. Surely they must also do something generalist?

   Please, see our replies above and the paragraph dedicated to this topic in the Discussion.

   1. In most figures, the authors do not compare flies with shared genetic backgrounds.

   The MyD88 allele we are using is a transposon insertion from the Exelixis collection and we are using the wA5001 strain that was used to generate the collection of insertion (Thibault et al., Nat. Genetics 2004). We thank the reviewer for this comment as we realized we had forgotten to mention the Bom__D__55C strain. Lines 603-604 state that the deficiency line has been isogenized in the wA5001 background.

   The phenotypes are usually strong so I am not concerned.

   However the rescue effect of Bom transgenes in Fig 5C-D is based on smaller differences. Were these genetic backgrounds controlled?

   Yes, as much as we reasonably could. The fact that most BomS transgenes did not rescue gives further confidence in the data.

   Were transgenes inserted at the same site?

   We used the strategy for overexpression developed by the Basler laboratory (Bishof et al., Development 2013, Nat. Protocols 2014) that relies on insertions at the same site.

   The authors seemingly used a heat shock to express transgenes.

   Heat-shocks are usually a short exposure to higher temperatures, usually 37°C. Here, we have used the inducible Gal4-Gal80ts system developed by McGuire and Davis (Trends in Genetics, 2004). The Gal80 repressor inhibits Gal4 function at the permissive temperature (18°C) and becomes inactive at the restrictive temperature (29°C). Thus, we use a temperature shift and not a bona fide heat shock.

   Given a resilience effect is being studied, this heat stress approach is sub-optimal. Earlier experiments showing effect/no effect of Bomanin on heat shock resilience would improve confidence here. I would recommend assaying temperatures that can kill wild-type in order to confirm that Bom do not succumb earlier (ex. up to 37'C).

   The results have been discussed above and show that 29°C is not a concern for Bom__D__55C and not much of a significant problem as regards MyD88.

   In Fig5C the time resolution is poor, and the effect inconsistent across Bomanins. What are the differences in the Bomanins that the authors suspect could cause this? And how consistent are the experiments?

   We provide all the primary survival data in Rebuttal Fig.1 A-H. The partial protection effects of BomBc1, BomS3 and BomS6 against restrictocin are consistent in the three independent experiments (Fig. 5D and Rebuttal Fig. 1 A-B). As regards the seven independent experiments performed with verruculogen, we observed a strong protection conferred by BomS6 expression in six experiments whereas we detected a milder protection conferred by BomS1 in four out of seven experiments and no protection in the three other ones. The effects were always there after 24 hours, in keeping with our novel data showing that BomS6 expression allows a faster recovery, around 10 hours, from verruculogen-induced tremors (Fig. 6E-F).

   Since the effect is finished by 24h, perhaps a boxplot of percent survival at this time would better show the consistency across experiments.

   Given the argument presented just above and considering that this rebuttal letter will be published alongside the article, this may not be needed.

   Minor concerns:

   i) The authors say the fungal burden of Bom55C flies remains low in Fig 5B, but they never measure flies that are near death when fungal load is greatest, or FLUD like in other figures. Given low mortality at the following time points, it seems likely that A. fumigatus would grow beyond initial loads in those individuals and kill them. I grant that these loads are less than what is seen in Hayan mutants. I just might suggest a more careful consideration of the time points used and what can be said about the trends shown here.

   This is certainly a relevant point. The FLUD data are now presented in Fig. EV8_A_ and do not reveal any additional growth.

   ii) Could the authors comment somewhere about the levels of toxin they were required to inject to get a phenotype vs. the level of toxins the authors expect are found in the fly during infection? I appreciate that toxin injection likely requires much higher doses, but it would be good to know just how far the authors have pushed their experimental system beyond its natural range.

   This a question that is difficult to answer accurately as we are not sure the techniques exist to measure toxin levels in these small flies. We have tested a range of concentrations. It is clear that we push the system and likely use concentrations that are higher than those actually secreted by A. fumigatus during infection. Indeed, the mutant strains defective for the production of verruculogen or restrictocin display only a mildly reduced virulence in MyD88 flies. This makes it even more remarkable that wild-type flies are able to withstand these high, unphysiological concentrations, an argument for an indirect effect independent of the dose as pointed out now in the Discussion. How fungal pathogens balance the expression of the hundreds of secreted virulence factors, proteins and secondary metabolites, is a major frontier for future investigations be them plant or animal pathogenic fungi/

   Again regarding toxins vs. general stresses, one could manage to inject salt into the hemolymph and show a stress-sensitized fly would succumb at lower doses than wild-type, emphasizing the relevance of defining concentrations.

   We feel that just monitoring the survival of flies after a challenge that produces an effect is sufficient (Fig. EV8_C_).

   The authors could also write toxin concentrations clearly in the figure/legend per experiment.

   Corrected.

   iii) Throughout the manuscript, the order that figures/panels are cited is inconsistent. Perhaps the text could be re-written so the reader can follow the figures more intuitively while going through the text?

   Corrected.

   iv) There are a few points where run-on sentences, involving many commas, make it hard to follow the logic. I might suggest a careful reading to break up long sentences into two sentences to ensure clarity.

   We hope to have addressed this concern.

   v) Line 279-281: this is the first and only mention of the immune surveillance hypothesis in nematodes. This is strange, given the authors are effectively describing an analogous idea exists in flies? Perhaps this could be added somewhere in the introduction or discussion.

   We have followed the advice of the reviewers and now discuss this point more fully in the Discussion under its own subheading.

   Small points

   • What timepoints are the gene expression data from? Could the authors indicate this in figures/legends?

   Done

   • Line 133-135: "We conclude that MyD88 flies succumb to a low A. fumigatus burden..." - could the authors cite a figure panel here to emphasize what evidence they're referring to.

   Done

   • Line 151-152: Dudzic et al. (2019- Cell Reports Figure 3) showed that PPO2 was regulated by Hayan, while PPO1 by Sp7. This relevant study should be cited here or in the introduction/discussion.

   Excellent suggestion, this was indeed an important study. Done

   • Line 179-180: could the authors define the gliotoxin mutant strain here in the text for clarity?

   Done

   • Line 196: Fig. 4A-B should be Fig. **S4 A-B?

   Corrected.

   • Fig4A: perhaps the authors could reduce the x-axis to focus on the early time points? If I understand correctly, aspf1 has slightly delayed killing compared to akuB (˜50% difference at 2 days), but both kill 100% by 3 days.

   Done

   • Fig4G: can the authors define the GFP transgene on pg10? Not clear what this is, or what this means. Brain? Fat body? The legend of Fig4G and the key in the top left... it's not easy to quickly understand what is shown in Fig4G.

   Done

   • Line 247: I would drop the "at the intracellular level" part. I'm not sure this is robustly shown given the use of an in vitro model where there is no closed extracellular environment. The data are convincing of the effect, this is just a semantic point.

   We agree that there is no closed extracellular environment and that therefore any signal emitted by the cells might get too diluted. However, the addition of EGF will activate the Toll intracellular through the chimeric EGFR-Toll receptor. As restrictocin is known to act intracellularly, one might have though that there might be some intracellular effectors mediating the Toll-dependent protection against restrictocin. Our sentence excludes this possibility.

   • Line 257-258: Cohen et al. (2020- Front Imm) never used Bomanin mutants. Did the authors mean to cite Hanson et al. (2019) here, which seems to fit their described citation re: Bom55C vs. Toll mutant flies (Fig. 2)? Given Hanson et al. infected Toll mutant and Bom55C flies with many bacteria/fungi including A. fumigatus, it's strange this study is not discussed currently.

   The reviewer is correct. Cohen et al. did use A. fumigatus, but on Daisho mutants and MyD88 and not Bom__D__55C as a control. We are now citing Hanson et al., 2019 in lines 443-449 (Discussion).

   • Fig5C-D: the labeling is difficult to follow.

   This is difficult to address unless multiplying EV figures. We feel this is not needed: the important curves are in color and each such curve is seen on the graphs.

   • Line 318: a -possible- AMP role of Bomanins was proposed because of the aforementioned killing activity of wt but not Bom mutant hemolymph, alongside rescue by single Bom genes. To say this was based only on survival experiments is incorrect.

   The paragraph has been rewritten and expanded to dispel any misunderstanding.

   • Line 324-328: could the authors cite appropriate references after "inhibition of calcium-activated K+ channels" ?

   Done

   • comment re-Line 334: Toll10b flies have melanotic tumors and are in general in a stressed state. Might their rescue be due to increased stress tolerance by pre-activated stress responses?

   This is a developmental effect occurring during larval stages, also observed for Cactus mutants. Here, we use a UAS-Tl10B transgene that is induced only at the adult stage using the Gal4-Gal80ts system. Thus, any stress is minimized as much as possible. Furthermore, we can phenocopy this phenotype to a large extent using a UAS-BomS6 driver, even though the phenotypes are subtly different as regards the protection against verruculogen-induced tremors.

   Referees cross-commenting

   Yes I agree that the data themselves are not the issue, nor even the direction of the results. But there are many overly-strong statements that go so far as to refute ideas which are supported by other studies, and for which the authors here do not provide any contradictory evidence.

   We hope that this revised, extended version has clarified any misunderstanding in the initial version.

   As per my review, I would be happy with a re-write that softened the language overall. I genuinely wonder if these Bomanin mutants simply have poor development, and so they are susceptible to neurotoxins/stress because their nervous system/development leaves them less resilient in general. Experiments testing their resilience to different stresses would greatly elevate the ability to make confident insights in the present manuscript. Currently the authors have only investigated one type of phenotype and interpreted it as if that is evidence of the evolved purpose of the peptides. This approach does not account for many other possible (and reasonable) explanations.

   We have performed the experiments suggested by the reviewer. While we see a modest effect of heat on MyD88, it is not found in Bom__D55C flies, which display essentially the same phenotype as MyD88 with regards to the sensitivity to A. fumigatus or some of its secreted mycotoxins_._

   Reviewer #2 (Significance):

   This paper should be of broad interest to the study of immunology, where roles for effectors are typically thought of as cytokines. In fruit flies and other invertebrates that lack adaptive immunity, immune effectors are more thought of as direct actors likely with antimicrobial properties. The finding that Toll might mediate resilience is interesting, and implicating well known Toll effectors provides an important step forward towards a mechanistic basis behind this resilience effect.

   We thank the reviewer for his appraisal of the significance of our work.

   My expertise is in insect and innate immunity.

   Read the original source
2. 

   Review Commons

   Nov 17, 2022

   Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Referee #2

   Evidence, reproducibility and clarity

   Summary:

   Xu et al. describe how A. fumigatus kills Toll-deficient fruit flies not by hyperproliferation, but more likely by virulence factors. Melanization is important for suppressing fungal spread. The Bomanin genes have an unknown function, and here the data suggest a reasonably convincing role for Toll in resilience. Overall the manuscript is thorough and presents a diversity of approaches that show Toll and the Bomanins in particular contribute to this resilience effect. The idea that Toll effectors are essential for resilience is interesting as other fly stress response pathways like JAK-STAT are better known for helping the …

   Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Referee #2

   Evidence, reproducibility and clarity

   Summary:

   Xu et al. describe how A. fumigatus kills Toll-deficient fruit flies not by hyperproliferation, but more likely by virulence factors. Melanization is important for suppressing fungal spread. The Bomanin genes have an unknown function, and here the data suggest a reasonably convincing role for Toll in resilience. Overall the manuscript is thorough and presents a diversity of approaches that show Toll and the Bomanins in particular contribute to this resilience effect. The idea that Toll effectors are essential for resilience is interesting as other fly stress response pathways like JAK-STAT are better known for helping the fly cope with damages, while Toll is better known as an antifungal response.

   I believe the study, with some careful considerations added, would add a valuable series of observations to understanding how the host immune system promotes survival after infection. Overall I am quite positive about the results, and the authors have made a significant effort. Any experiment suggestions I make are strictly to improve the confidence in the interpretations of the results, but the language could alternately be softened to address those concerns. My major critique is that the authors repeatedly extend beyond what is shown, and occasionally in defiance of what is shown (if I understand the results correctly). Comments below.

   Major comments:

   1. The language is too strong. Specifically the use of the phrase "anti-toxin" is too generalist, especially as the authors show that their candidate Bomanin does not bind to the toxin directly. Instead, Toll mutants seem susceptible to damage/stress caused by injury/toxins. MyD88 even show general susceptibility to vehicle controls in Fig3C-D. Toll is important for development, so it may be expected that Toll flies could have development defects impacting resilience even if/when Toll flies can survive to adulthood. I don't say this to be too negative on the findings, which are quite convincing. But I am not sure that the phrase "anti-toxin" is right for what is shown.
      A very interesting recent study shows Dif has a role in the synapse of neurons to protect from alcohol sensitivity. Could secreted Bomanins participate? This emphasizes a mechanism through which Toll mutants likely have defective neural development, which could make them stress response defective, especially to things like neurotoxins. See: https://pubmed.ncbi.nlm.nih.gov/35273084/
      Lin et al. (2019) also showed lack of Bomanin secretion from the fat body in Bombardier mutants causes loss of tolerance (resilience?). So does Bomanin disruption increase susceptibility to stresses more generally, rather than specifically fungal toxins? And is this a development role, rather than an immune response role?
      The authors could try to use other stresses (NaCl, oxygen, heat, alcohol) to test the contribution of Bomanins to this resilience, which may reflect defective neural development rather than a role for secreted systemic immune-response peptides.
   2. The authors present a paradox. On the one hand, A. fumigatus hardly induces Drs/Bomanins (Fig. S1). Yet on the other, they propose that inducible Bomanins protect the fly from mycotoxins. Why do the authors say Toll is hardly induced by A. fumigatus at the start of the study (Fig S1), but later use the same data to argue that Bomanin induction underlies the resilience phenotype (Fig5). In Fig 5, it looks like DMSO is nearly identical to A. fumigatus, so can the authors really suggest that equal induction to DMSO is relevant?
      The authors' discussion of these points would benefit from considering Vaz et al. (2019; Cell Rep) to frame how much PAMP is injected given equal numbers of fungal cells vs. bacterial cells. To me the lower induction by injecting a few fungal cells with much lower surface area to volume ratio means equal microbe mass has exponentially less PAMP in fungal conidia cell walls (2-3um diameter) vs. equal mass of bacteria (0.5-1um diameter).
      Fig S1O is not convincing that Boms alone are present. There is significant noise near Drs in FigS1 infected, which likely saturates the detector before Drs can fly to it. I say this because DIM4 (Daisho) indicates that Toll is strongly induced. The authors should show a larger mass range on the x-axis including peaks of other Toll-induced peptides like the BaramicinA DIM10, DIM12 and DIM13 peptides of their companion paper and DIM14 (Daisho), which are closer in mass to the Bomanins and less likely disrupted by the noise at 4300 m/z. The maldi-tof calibration to correct ranges is critical for arguments of quantification.
      Combined with comments in Major Concern 1, I am not convinced that the -inducible- Bomanin response mediates the resilience phenotype.
   3. The author's language is very strong to disregard a possible antimicrobial activity. Previous studies showed increased Candida growth and decreased hemolymph killing activity in Bom55C flies (Lindsay et al. 2018 and Hanson et al. 2019). Also see minor concern (i).
      I grant that the data are consistent with a resilience role. However the authors found no binding of Bomanin to restrictocin, countering their idea of a -direct- anti-toxin effect. At present the authors cannot rule out a direct antimicrobial role, or even the possibility of two different roles for the same peptides (ex: one in resilience, one antimicrobial). For instance, it is difficult to explain the loss of killing activity of Bom-deficient hemolymph ex vivo from Lindsay et al. if Bomanins are strictly anti-toxins. Surely they must also do something generalist?
   4. In most figures, the authors do not compare flies with shared genetic backgrounds. The phenotypes are usually strong so I am not concerned.
      However the rescue effect of Bom transgenes in Fig 5C-D is based on smaller differences. Were these genetic backgrounds controlled? Were transgenes inserted at the same site? The authors seemingly used a heat shock to express transgenes. Given a resilience effect is being studied, this heat stress approach is sub-optimal. Earlier experiments showing effect/no effect of Bomanin on heat shock resilience would improve confidence here. I would recommend assaying temperatures that can kill wild-type in order to confirm that Bom do not succumb earlier (ex. up to 37'C).
      In Fig5C the time resolution is poor, and the effect inconsistent across Bomanins. What are the differences in the Bomanins that the authors suspect could cause this? And how consistent are the experiments? Since the effect is finished by 24h, perhaps a boxplot of percent survival at this time would better show the consistency across experiments.

   Minor concerns:

   • i) The authors say the fungal burden of Bom55C flies remains low in Fig 5B, but they never measure flies that are near death when fungal load is greatest, or FLUD like in other figures. Given low mortality at the following time points, it seems likely that A. fumigatus would grow beyond initial loads in those individuals and kill them. I grant that these loads are less than what is seen in Hayan mutants. I just might suggest a more careful consideration of the time points used and what can be said about the trends shown here.
   • ii) Could the authors comment somewhere about the levels of toxin they were required to inject to get a phenotype vs. the level of toxins the authors expect are found in the fly during infection? I appreciate that toxin injection likely requires much higher doses, but it would be good to know just how far the authors have pushed their experimental system beyond its natural range. Again regarding toxins vs. general stresses, one could manage to inject salt into the hemolymph and show a stress-sensitized fly would succumb at lower doses than wild-type, emphasizing the relevance of defining concentrations. The authors could also write toxin concentrations clearly in the figure/legend per experiment.
   • iii) Throughout the manuscript, the order that figures/panels are cited is inconsistent. Perhaps the text could be re-written so the reader can follow the figures more intuitively while going through the text?
   • iv) There are a few points where run-on sentences, involving many commas, make it hard to follow the logic. I might suggest a careful reading to break up long sentences into two sentences to ensure clarity.
   • v) Line 279-281: this is the first and only mention of the immune surveillance hypothesis in nematodes. This is strange, given the authors are effectively describing an analogous idea exists in flies? Perhaps this could be added somewhere in the introduction or discussion.

   Small points

   • What timepoints are the gene expression data from? Could the authors indicate this in figures/legends?
   • Line 133-135: "We conclude that MyD88 flies succumb to a low A. fumigatus burden..." - could the authors cite a figure panel here to emphasize what evidence they're referring to.
   • Line 151-152: Dudzic et al. (2019- Cell Reports Figure 3) showed that PPO2 was regulated by Hayan, while PPO1 by Sp7. This relevant study should be cited here or in the introduction/discussion.
   • Line 179-180: could the authors define the gliotoxin mutant strain here in the text for clarity?
   • Line 196: Fig. 4A-B should be Fig. **S4 A-B?
   • Fig4A: perhaps the authors could reduce the x-axis to focus on the early time points? If I understand correctly, aspf1 has slightly delayed killing compared to akuB (˜50% difference at 2 days), but both kill 100% by 3 days.
   • Fig4G: can the authors define the GFP transgene on pg10? Not clear what this is, or what this means. Brain? Fat body? The legend of Fig4G and the key in the top left... it's not easy to quickly understand what is shown in Fig4G.
   • Line 247: I would drop the "at the intracellular level" part. I'm not sure this is robustly shown given the use of an in vitro model where there is no closed extracellular environment. The data are convincing of the effect, this is just a semantic point.
   • Line 257-258: Cohen et al. (2020- Front Imm) never used Bomanin mutants. Did the authors mean to cite Hanson et al. (2019) here, which seems to fit their described citation re: Bom55C vs. Toll mutant flies (Fig. 2)? Given Hanson et al. infected Toll mutant and Bom55C flies with many bacteria/fungi including A. fumigatus, it's strange this study is not discussed currently.
   • Fig5C-D: the labeling is difficult to follow.
   • Line 318: a -possible- AMP role of Bomanins was proposed because of the aforementioned killing activity of wt but not Bom mutant hemolymph, alongside rescue by single Bom genes. To say this was based only on survival experiments is incorrect.
   • Line 324-328: could the authors cite appropriate references after "inhibition of calcium-activated K+ channels" ?
   • comment re-Line 334: Toll10b flies have melanotic tumors and are in general in a stressed state. Might their rescue be due to increased stress tolerance by pre-activated stress responses?

   Referees cross-commenting

   Yes I agree that the data themselves are not the issue, nor even the direction of the results. But there are many overly-strong statements that go so far as to refute ideas which are supported by other studies, and for which the authors here do not provide any contradictory evidence.

   As per my review, I would be happy with a re-write that softened the language overall. I genuinely wonder if these Bomanin mutants simply have poor development, and so they are susceptible to neurotoxins/stress because their nervous system/development leaves them less resilient in general. Experiments testing their resilience to different stresses would greatly elevate the ability to make confident insights in the present manuscript. Currently the authors have only investigated one type of phenotype and interpreted it as if that is evidence of the evolved purpose of the peptides. This approach does not account for many other possible (and reasonable) explanations.

   Significance

   This paper should be of broad interest to the study of immunology, where roles for effectors are typically thought of as cytokines. In fruit flies and other invertebrates that lack adaptive immunity, immune effectors are more thought of as direct actors likely with antimicrobial properties. The finding that Toll might mediate resilience is interesting, and implicating well known Toll effectors provides an important step forward towards a mechanistic basis behind this resilience effect.

   My expertise is in insect and innate immunity.

   Read the original source
3. 

   Review Commons

   Nov 17, 2022

   Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Referee #1

   Evidence, reproducibility and clarity

   Summary

   The authors have set out to study the Drosophila immune response against the fungus Aspergillus fumigatus. They found that Aspergillus fumigatus kills Drosophila Toll pathway mutants. The fungus does this without invasion because its dissemination is blocked by melanization. They suggest that there is a role for Toll in host defense distinct from resistance. The findings are interesting, and looks like the mycotoxins play a role. It also seems that there is some role of the Bomanins here, but I find that in particular Figure4 experiments are not convincing enough to provide a mechanistic insight as to what is going on. I …

   Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

   Learn more at Review Commons

   ---

   Referee #1

   Evidence, reproducibility and clarity

   Summary

   The authors have set out to study the Drosophila immune response against the fungus Aspergillus fumigatus. They found that Aspergillus fumigatus kills Drosophila Toll pathway mutants. The fungus does this without invasion because its dissemination is blocked by melanization. They suggest that there is a role for Toll in host defense distinct from resistance. The findings are interesting, and looks like the mycotoxins play a role. It also seems that there is some role of the Bomanins here, but I find that in particular Figure4 experiments are not convincing enough to provide a mechanistic insight as to what is going on. I think the authors need to think through what their results mean, and also, explain better (especially regarding Fig 4) their ideas and how the data fits them.

   Major comments:

   Page 5: .."the fungal burden did not increase much in MyD88 flies challenged with 50 conidia (Fig. 1B)" - What do you mean did not increase much? There is a clear increase in Myd88 mutants compared to controls; would you expect a bigger increase (e.g. log scale induction)? Explain.

   Page 6: "the SPZ/Toll/MyD88 cassette is required for host defense against A. fumigatus infections, even though this pathogen only mildly stimulates the Toll pathway." - Should you rather say that A. fumigatus only mildly induces the Toll pathway target gene Drosomycin?

   Page 6: "...we tested Hayan mutant flies defective for this arm of innate immunity (Nam et al., 2012)." - elaborate this, which arm/which pathway? This has not been introduced in the introduction. Explain. Can you really draw this conclusion: "We conclude that melanization limits the proliferation and the dissemination of A. fumigatus injected into wild-type flies yet does not eradicate it at the injection site, where a melanization plug forms." Maybe you can based on the function/importance of the pathway to melanization, but you need to explain.

   Page 10: "The cleavage of the 18S RNA was however much less pronounced in wild-type flies as compared to MyD88" - I am not sure what this means. Do you mean 28S? And that the 28S peak is lower? Is this a quantitative method? Fig. legend: "Arrows show the position of the 28S RNA sarcin fragment" - there are three arrows in both Fig 4E and F; specify which arrows point what.
   Based on the results, I am not convinced about the conclusion, that "restrictocin is able to inhibit translation to a detectable degree in vivo, likely through the cleavage of the ribosomal 28S a-sarcin/ricin loop as described in vitro." <- Do you draw this conclusion before doing the actual in vitro experiment, which is described next in the text (The rabbit reticulocute assay, S2 cells)?

   4H: Not sure what should be seen here, is it the darkest band at 0 uM that disappears? HI & J need more explanation than what is now included in the text or Figure legend, is the conclusion that there is no difference? Write the stats above the Figs 4I & J (n.s.?).

   Minor comments:

   It would have helped commenting if the manuscript contained line numbers

   Why do you have the title "Hayan" on top of Fig 1F; you don't have this marking system in the other survival curves

   Fig 2A: Can you speculate why MyD88 flies die rapidly at day 10 if you inject PBST (your control)? What would happen to uninjected controls in otherwise the same conditions? (you could include an uninjected control here?)

   Fig 2E: Not sure what would be the best way of presenting the curves - different colors, dotted lines or something? Now if there are too many lines, they are hard to tell apart. because the symbols are not that visible. Like in 2E if you want to compare the light red/orange colored lines.

   For consistency add the caption also to Fig 3D (I assume it is the same as 3C)

   For consistency, should you add Verruculogen on top of Fig 3F?

   Chronologically, how it is explained in the text, Figs 4A and B are in the wrong order.

   The quality of Fig 4 is not great, the text is hard to read (too small) and becomes blurry upon magnification.

   Page 12; "These data then suggest that a process akin to the immune surveillance of core cellular processes first described in C. elegans may also exist in Drosophila" - I think this sentence belongs to the discussion, this is not directly drawn from the results.

   Referees cross-commenting

   I think we share many thoughts among all the reviewers. The main problem is that the manuscript language is quite strong; from the results many times it is not ok to make such strong statements. Some experiments need further analysis and clarification. I think in most cases, this could be achieved by softening the statements and adding more discussion, and not by making new experiments (some may be needed).

   Minor things are that experiments are not advancing in a logical order between the text and the figures and there are problems with resolution in some figures. Statistics in some figures needs to be added.

   Significance

   The nature of the work is conceptual for the field, to understand the role of the Toll pathway and Bomanins in particular, in this fungal infection model. The work is interesting to a somewhat limited audience, mainly immunologists and in particular, people interested in the Drosophila model for immunity. The work may be interesting conceptually in understanding fungal infections.

   My expertise: I am a Drosophila immunity researcher with nearly 20 years of experience in working with fly immunity, in particular the Toll and the Imd pathways.

   Read the original source
4. 

   Version published to 10.1101/2022.08.18.504437v1 on bioRxiv

   Aug 23, 2022